Description:
<jats:title>ABSTRACT</jats:title><jats:p>Near‐isogenic line (NIL) libraries can be used to detect beneficial trait variation in germplasm that is unadapted or has poor agronomic performance. The objectives of our study were to compare the <jats:italic>t</jats:italic> test, Dunnett test, and linear model test with regard to the power and false positive rate of quantitative trait loci (QTL) detection in NIL libraries of different design. We employed computer simulations with maize genome models to investigate nonoverlapping NIL libraries, overlapping NIL libraries, and stepped aligned inbred recombinant strains (STAIRS) libraries for traits with oligogenic inheritance. Quantitative trait loci detection power of the linear model and Dunnett tests were similar for nonoverlapping and STAIRS libraries; for overlapping NIL libraries the Dunnett test was slightly superior. False positives were greatest for the <jats:italic>t</jats:italic> test and lowest for the linear model test. False positive sums with the Dunnett test were generally higher than for the linear model test if the heritability was 0.9 or lower. We conclude that the linear model test is superior to the Dunnett test for nonoverlapping NIL libraries and for overlapping NIL libraries with heritabilities below 0.9, as usually occur. Analysis of a rapeseed (<jats:italic>Brassica napus</jats:italic> L.) library revealed two other major advantages of the linear model test. First, detection of positive and negative QTL effects present in the same line is possible. Second, for NILs with multiple donor segments, observed phenotypic differences can be assigned to individual chromosome segments.</jats:p>